3-31-2021
A Pandemic of Protectionism: How Economic Isolationism
A Pandemic of Protectionism: How Economic Isolationism
Affects the Economy
Affects the Economy
Álvaro MartínUniversity of Cambridge, am2875@cam.ac.uk
Follow this and additional works at: https://jnf.ufm.edu/journal
Part of the International Economics Commons, Macroeconomics Commons, and the Political Economy Commons
Recommended Citation Recommended Citation
Martín, Álvaro (2021) "A Pandemic of Protectionism: How Economic Isolationism Affects the Economy," Journal of New Finance: Vol. 2 : No. 1 , Article 1.
DOI: 10.46671/2521-2486.1011
The aim of this paper is to study the economic consequences of protectionism from a macroeconomic perspective. For that purpose, we estimate the impact of tariffs on different economic variables as GDP growth, trade volume or unemployment, comparing the effects for two groups of countries: high-income and low-income. For this, we build a regression model to estimate the effects of tariff changes on each of these variables, differentiating between the two groups of countries. Tariffs contribute to increase
unemployment, generate lower economic growth and reduce trade volumes, but by how much? This is very relevant in the actual pandemic environment in terms of looking for efficient public policy for
economic recovery. This is complemented with an overview of protectionism and trade restrictions in the pandemic environment, and how economic nationalism could affect policymaking in the recovery phase of the actual recession. Multilateralism seems to be at risk due to these political and economic trends, with multilateral organizations having had their role as trade arbiter eroded, which could be a threat to global value chains and free-trade mechanisms. We conclude that recent tariffs have no significant effect on high-income countries, while variables as GDP growth rates or unemployment in low-income countries were more sensitive to variations of tariff rates.
Keywords Keywords
Macroeconomics; Multilateralism; Protectionism; Tariffs; Trade.
JEL Code F10, F13, F14, F16, F62 Submission Date 3-3-2021 Approval Date 3-10-2021 Publication Date 3-31-2021
1. Introduction
Since Adam Smith (1776), writing in the era of mercantilism, countries use of their competitive advantages in trade have been extensively debated. Since the consolidation of modern economics, there has been general agreement among economic practitioners that free trade and free markets are the best possible option for allocating resources in an efficient way, generating the least distortions and maximizing output by directing resources to their most productive employment. There is also general agreement about the existence of market failures and imperfections, but tariffs have been shown to be a suboptimal solution for these problems. Tariffs lead to a less productive allocation of resources and promote economic inefficiencies which end up generating greater overall losses to consumers than gains for producers, without even considering retaliation and other collateral effects of tariffs. The main channel by which import tariffs reduce welfare is by introducing a wedge between the marginal social cost and marginal social benefit of the imported goods. This doesn’t tend to be the only effect of tariffs, as their distortionary nature affects industries differently and widens the productivity and efficiency gaps between them.
Most analysis regarding the effects of tariffs on the economy has been of a microeconomic nature, analyzing industry-level effects (Grossman and Rogoff, 1995). These analyses are of great value, but tariffs in the modern economy have proved to have economy-wide effects, with trade policy becoming both a macroeconomic and a geopolitical policy tool to be used by governments. Studying the macroeconomic effects of tariffs has become crucial nowadays, as protectionism has made its comeback since the beginning of the second decade of the century, intensifying with the eruption of coronavirus and its effects on global supply
Protectionist policies, such as the Smoot-Hawley tariff, were one of the main causes of the international trade collapse in the 1930s. Trade took off again after the Second World War, promoted by the liberal-democratic international world order on which Western countries agreed after 1945. In this paper, we carry out an empirical study1 of the relationship between
tariff barriers and the behavior of relevant macroeconomic variables such as GDP growth, unemployment or trade volume. To do so, I set up a regression model based on Jorda’s (2005) local projection method. This methodology allows us to account for non-linearities without the need to impose strict and unnecessary dynamic restrictions. For this purpose, I use data of high-income and low-high-income countries (as classified by the World Bank, 2021) trying to estimate the differential effect of tariffs on both groups. Therefore, I analyze how variations in tariff rates relate to changes in the aforementioned economic variables, analyzing the diverging effects for the two groups of countries. Afterwards, I study how the new protectionist wave could worsen recovery from the Covid-19 recession by disrupting multilateral trade structures and their main multilateral organizations such as the WTO. However, high uncertainty levels make predictions regarding the future of trade difficult.
Even though I use an extensive dataset, it is nearly impossible to reliably control for the effect of structural policies in relation to variations of tariff rates, which could mitigate or worsen the effects of the latter. The cause for this is mainly the absence of data for the vast majority of the low-income countries analyzed, which only have available data concerning basic economic variables and their corresponding indicators. I use the wide span of data to build a solid regression model for the proposed analysis, always dependent on empirical validation and robustness checks for the model.
Section 2 of this paper presents the methodology, data sources and regression model. Section 3 introduces the baseline results segmented by the two groups of analyzed countries:
separately the effects of variations in tariff rates on GDP growth, variations in unemployment rates and trade volumes. In section 4 of the study, I analyze how the coronavirus pandemic has affected international trade relations and conclude by saying that the pandemic has only deepened multilateralism’s loss of popularity among world leaders, leading to an even greater protectionist backlash. Finally, I summarize the paper’s main conclusions regarding the effects of tariffs on certain economic variables and their implications for public policy to combat the coronavirus recession.
2. Methodology and data
This paper sets out to study the dynamic response of GDP, unemployment and trade to variations in tariff rates. For this purpose, a regression model is used which prevents unnecessary constraints, while employing a wide span of data for high-income and low-income countries. This model makes it possible to obtain a macroeconomic vision of the effects of tariffs, quantifying them for a five-year period, which is the average estimated response time of said economic variables in relation to changes in tariff rates. This study is based on data gathered up to 2017 because for some of the studied variables, it was the most recent registered data available on the IMF and World Bank databases, especially in the case of low-income countries. However, the regression model is valid for any timespan or group of countries. The regression model employed for this purpose is as follows:
𝑌𝑖𝑡+𝑗− 𝑌𝑖𝑡−𝑔 𝑦𝑖𝑡−𝑔
= 𝛼𝑖+ 𝛽𝛥𝑇𝑖𝑡 + 𝜀𝑖𝑡 + 𝜀𝑖𝑡(𝑌𝑡, 𝑊𝑉𝑖, 𝑡)
In this regression, 𝑌𝑖𝑡+𝑗−𝑌𝑖𝑡−𝑔
𝑦𝑖𝑡−𝑔 is the change in the outcome variable analyzed for each
or trade volumes. ∝𝑖 controls for unobserved country specific characteristics and their
heterogeneity, being different for each group of countries or studied timespan (IMF, 2019).
∆𝑇𝑖,𝑡 represents the change of the tariff rate. Finally, 𝜀𝑖,𝑡 stands for the error of the
regression, as a function of 𝑌𝑡, which stands for controls for global shocks and their effects
on both country groups and 𝑤𝑣𝑖,𝑡, which stands for the product of vectors of control variables
and their corresponding coefficients.
The data used for the empirical analysis is obtained from the World Economic Indicators (World Bank, 2021) and the World Economic Outlook (IMF, 2021), in relation to the output variables. However, for tariff rates and their intertemporal changes, data is obtained from the World Trade Organization database (WTO, 2021), which also provides data related to relative weights of imports for each industry. As tariff rates are normally provided by sector, I use Topalova and Khandewal’s (2011) computation for average tariff rates, where the average tariff is calculated as a weighted average of sectorial tariffs, with the corresponding weights being the share of imported inputs in each sector, which is a proxy for how sensitive each sector can be to tariffs, regarding variations in input costs. For this, I employ the following weighted average formula:
𝑇𝑠, 𝑖, 𝑡 = ∑ 𝛿𝑠, 𝑖, 𝑡𝑇𝑘, 𝑖, 𝑡
𝑘
For this formula, 𝑇𝑠,𝑖,𝑡 is the respective tariff rate applied for sector S, for which 𝛿
3. Baseline results
3.1. Tariff rate variation and GDP growth
Firstly, we proceed to investigate the coefficient of determination and the regression coefficient between changes in tariff rates and GDP growth for high income countries, even though a full causality relation cannot be obtained from this study. This is due to omitted variable biases, arising from lack of data to control for sufficient variables to prevent this bias. However, the coefficient of determination and of regression should give us a picture of the relationship between these two variables. It is important to note that statistical and economic significance of results are not interchangeable. As it will be observed throughout the analysis, the relationship is not direct.
To obtain a clearer picture we need to look carefully at the results obtained from the regression and correlation analysis.
By observing the results from the correlation and regression analysis (below), we can see that the 𝛽 obtained is -0.097742, showing that for a 1% increase in tariff rates, GDP growth rate would be reduced by 0.01% in the case of high-income countries, which does not look like a significant direct relationship. The coefficient of determination is also weak in this case, being just 0.0377, showing a significantly weak correlation for the observed data. To check for the robustness of our results, we look for p-values, of 0.713, which represent no evidence against the tested null hypothesis of 𝛽=0, for a 5% significance level.
We now repeat the test for the case of low-income countries, for which we obtain significantly different results, as can be appreciated from the connected scatter plot below. Even though some of the data points present dispersion, we see a stronger linear relation than for the case of high-income countries. From this we can presuppose that there exists a stronger relationship between changes in tariff rates and GDP growth for low-income countries. However, this needs to be corroborated by carefully analyzing the results obtained from the correlation and regression analysis.
When looking at the results from the regression and correlation analysis (below), we can see that the 𝛽 obtained is -0.150981, showing that for a 1% increase in tariff rates, GDP growth in low-income countries would be reduced by -0.150981, which translates into a reduction of -1.5% of GDP growth when the increase in tariff rates oscillates around 10%: not unusual in low-income countries. The coefficient of determination is moderately strong in this case at 0.5327, representing a significant correlation between the observed data points. Economically, this result is very significant, as it shows that variation of tariff rates can explain 53% of variation in GDP growth rates in low-income countries, showing how economic policy regarding tariff rates can have large effects on the economy. To check for the robustness of our results, we look for p-values, of 0.1, which represent weak evidence against the tested null hypothesis of 𝛽=0, for a 5% significance level.
3.2. Tariff rate variation and change in unemployment rate
We now proceed to study how, if at all, tariff rate changes affect unemployment rates, for both high-income countries and low-income countries. To do so, we use the same methodology as before, studying high-income and low-income countries as separate samples of data. As before, a full causality relationship cannot be obtained from this study due to omitted variable biases, arising from difficulties regarding availability of data to control for sufficient variables to prevent this bias. However, the coefficient of determination and of regression should give us a picture of the relationship between the variables studied.
We start by analyzing the relationship between changes in tariff rates and variations in unemployment figure for high income countries. By looking at the connected scatter plot, we can see notable dispersion of data points, and once again no single linear relationship for the plotted data. The dispersion of data points enlarges when the tariff rate change is larger, while it remains constant for smaller changes in tariff rates.
From this, we can presuppose that a very weak relationship exists between changes in tariff rates and variations of unemployment levels for high income countries. We now proceed to corroborate these observations with a more in-detail analysis of the results from the correlation and regression analysis.
When looking at the results from the correlation and regression analysis (below), we observe that the 𝛽 obtained for this regression is -0.035942, which means that for each 1% increase in tariff rates in high income countries, GDP growth is reduced by -0.035942, which is not significant for our model. When looking for the coefficient of determination, we obtain a value of 0.0072, which represents very weak correlation between the studied variables. To check for the robustness of our results, we look for p-values, of 0.873, which represent no evidence against the tested null hypothesis of 𝛽=0, for a 5% significance level. These results preliminarily show that there is no strong relationship between changes in tariff rates and variations in unemployment levels in high-income countries.
We now repeat the same study for low-income countries, looking for the relationship between changes in tariff rates and variation in unemployment levels. For low-income countries, the connected scatter plot looks different when compared to that of high-income countries. We see a general linear relationship between the two studied variables, which would be stronger and clearer if it wasn’t for the outlying data point observed in the graph. This can be corroborated by proceeding to a more in-depth analysis of the results obtained from the correlation and regression study for the relation between changes in tariff rates and changes in unemployment levels for low-income countries.
When looking at the results from the correlation and regression analysis (below), we observe that the 𝛽 value obtained is 0.0736152, which is more significant than the one previously obtained for the case of high-income countries. This means that for low-income countries, where a 10% change in tariff rates is not an unusual phenomenon, this variation in tariff rates will cause a 0.74% increase in the rate of change of unemployment. However, when looking for the value of the coefficient of determination, we obtained a value of 0.3293, which represents a weak correlation between both variables, showing that changes in tariff rates account for just 33% of changes in unemployment rates in the case of low-income countries. To check for the robustness of our results, we look for p-values of 0.234 which represent no evidence against the tested null hypothesis of 𝛽=0, for a 5% significance level.
3.3. Tariff rate variation and change in volume of imports of goods and services We now turn to our final relationship in this study: the relationship between changes in tariff rates and changes in the volume of imports of goods and service in aggregate. The study is carried out separately for high-income and low-income countries, as throughout the rest of the paper. Also, as explained above, no absolute causality relationship can be obtained from the regression and correlation analysis without committing to an omitted variable bias, due to lack of available data, especially for low-income countries regarding other variables that could
affect our results. We will concentrate on studying the conclusions that can be obtained by analyzing the regression and determination coefficients.
Firstly, we analyze the relationship between changes in tariff rates and changes in the aggregated value of imported goods and services for the set of high-income countries. By looking at the connected scatter plot we can observe a notable dispersion of data points, with partial linear relationships that display contrarian trends. This does not provide a clear image of any existing relationship between the two variables. To obtain clearer conclusions, we will now proceed to analyze the results obtained from the correlation and regression analysis in greater depth.
By looking at the table which presents the results of the correlation and regression analysis (below), we can see that the value obtained for 𝛽 is -0.027964, which means that an increase in tariff rates of 1% will cause a variation in the change in volume of total imports in high-income countries of -0.03%, which is not really significant for the purposes of this study.
showing that just 4.7% of the change in volume of total imports can be explained by the variation rates of tariffs: a very weak coefficient of determination. To check for the robustness of our results, we look for p-values of 0.681, which represent no evidence against the tested null hypothesis of 𝛽=0, for a 5% significance level.
We now proceed to repeat the same test for the low-income set of countries. By observing the connected scatter plot, we see a similar image to that of high-income countries, where a high dispersion of data points exists and there is no general linear relationship. At the same time we see contrarian partial linear relationships between some of the data points, leaving a very unclear image regarding the relationship between the two variables under study. Let’s analyze the results obtained from the correlation and regression analysis in more depth.
When looking at the results from the correlation and regression analysis (below), we obtain a value for 𝛽 of -0.0692616. In the case of low-income countries, where oscillations in tariff rates can be even higher than 10% a year, an increase of 10% in the tariff rate could cause the variation of total volume of imports of goods and services to vary by -0.7%. When looking for the determination coefficient, we obtain a value of 0.0242, which is very weak, as it reflects that just 2.4% of the variation in the volume of imports could be preliminarily explained by changes in tariff rates in low-income countries. To check for the robustness of our results, we look for p-values of 0.769, which represent no evidence against the tested null hypothesis of
4. How has Covid-19 affected protectionist tendencies and
international trade?
In this section of the paper, we draw briefly on the possible effects of Covid-19 on protectionist tendencies and future trade policies, building on previous very good work that has been done on this subject.
A possibility that has been widely discussed over the last few months has been that Covid-19 exacerbates a protectionist political response following the economic effects of the pandemic, as discussed in the introduction to this paper. This protectionist response would have considerable effects on global value chains, which cannot be understood without taking a brief look at history, and particularly at how governments and global multilateral trade structures responded to previous recessions. The Great Depression of the 1930s acted as a trigger of tariff increases and export quotas, restricting international trade and reducing global mobility of capital. In the 1980s, voluntary export restraints followed the recessions of the time, and finally, the Great Recession starting in 2008 gave a new perspective to international trade. After 2008, instead of imposing greater tariffs and quotas, and consequently restricting global trade, governments all around the world decided that industrial policy would consist in greater subsidization of strategic industries, mostly manufacturing companies, national conglomerates
or agricultural lobbies. These types of policies haven’t disappeared since then, and even before the irruption of Covid-19, industrial policy was one of the main points over which China and the USA confronted each other in their much publicized trade war. Everything seems to indicate that after the Covid-19 crisis, economic nationalism will expand its dominions and protectionism will involve greater subsidization for national companies along with higher barriers to entry to national provision of goods and services.
Depending on the duration of the pandemic, governments will design different strategies to bail out certain companies or strategic sectors, which could go against regulation -such as that of the European Union- regarding competition.
But even if international trade, , was growing, it was already suffering even before Covid-19. For example, industrial costs since 2015 have been continuously rising in the US mainly due to interruptions in supply-chains and higher tariff-induced production costs (Albertoni and Wise, 2021).
In 2020, the first supply chain shock came from a strict lockdown in Hubei province, with multiple Chinese intermediate goods seeing their supplies collapse and thousands of companies all around the world having to paralyze their production. In response to these events, academics such as Javorcik (2020) spoke up about the urgent need for firm supply-diversification and a rethinking of their global value chains. However, this doesn’t entail the necessity of governments having to arrange trade or production structures, as some have interpreted. Global trade needs to be dealt with as a multilateral phenomenon, so national political autarchical responses will just put sand in the wheels of international trade without solving actual disputes, which need to be settled in a multilateral framework.
contributed to damaging an already sick multilateral trading system. This has made it even harder for the WTO to continue its efforts to deliver a multilateral round of trade negotiations to reduce global tariffs and restrictions to trade. The absence of the WTO from economic policy discussions throughout the pandemic has eroded its position as central arbiter of trade disputes. However, the WTO also entered the Covid-19 recession in a very difficult position as in December 2019 the Appellate Body ceased to execute its functions after the U.S. government blocked new judges’ appointments, leaving world trade orphaned in times of rising protectionism and continuous erosion of multilateralism and rules-based global dispute-settling mechanisms. The WTO has abdicated from its leadership in global trade matters, and this will give greater leeway to future protectionist and nationalist political movements to make their policies effective.
China constitutes another threat arising from the WTO’s abdication from its global rule. Having stopped its economically liberalizing reforms, China is strengthening its state capitalism model as described by Branko Milanovic (2019). As a result, China is imposing greater costs on other nations by providing illegal subsidies and privileges to public companies, and promoting a corporatist model with insufficient intellectual property protection for international firms and harsher restrictions to foreign direct investment, such as forced technological knowledge transfer requirements.
Furthermore, there are two main impacts of the Covid-19 crisis on international trade. This crisis is reinforcing previously existing trends, including the deceleration of growth of international trade volumes, the rise of economic insecurity following certain arbitrary trade-restrictive policies and the fallout of poor Sino-American relations. The pandemic, as already shown, has brought severe new challenges to international trade, with export protectionism rising and debilitating multilateral cooperation commitments. The second great impact of Covid-19 in terms of international trade perspectives is the danger of renationalization of
certain strategic companies, industries or sectors all around the world. Economic nationalism and the re-shoring of manufacturing have been deeply discussed by policymakers in Europe and the U.S. These policies will only contribute to the elimination of incentives to international cooperation and trade expansion, increasing input costs and eroding the competitiveness of many industries. The extension of global supply chains is the solution, not the problem. The aftermath of the Covid-19 recession will be a hard time for globalization, with major geopolitical threats and social disruption.
5. Conclusion
In this paper we have built a regression model to study the effects of tariff rate variations on the variation rates of dependent variables such as GDP growth, unemployment rates or the volume of aggregate imports. In each case we have developed a separate study for high-income and low-income countries, obtaining diverging results. In terms of the effect of tariff rate changes on the variation of GDP growth rates, we found no significant effect for high-income countries, while GDP growth rates for low-income countries were more sensitive to variations in tariff rates. This same pattern applied to the study of the relationship between variation in tariff rates and unemployment rate changes. In the case of high-income countries, therefore, the relationship was not significant, but, for low-income countries we found a slight relationship between the two variables, meaning that for a certain variation in tariffs, effects on GDP variation rates were greater than in high-income countries, showing a higher sensitivity of changes in GDP in low-income countries to variation of tariff rates. However, for the case of import volumes, we found no significant relationship with variation in tariff rates either for high-income or low-income countries. A preliminary conclusion - even causality cannot be fully assessed with the available data for low-income countries or the model used in this paper
and citizens suffer more from tariff restrictions to trade, due mainly to higher marginal cost or benefit from variation in tariff rates. Further research should be done in this area, with access to more detailed databases, especially for low-income countries. This would allow us to isolate certain parameters and study the causality between variation of tariff rates on variables such as GDP growth, unemployment levels or trade volumes.
In terms of the effects of Covid-19 on protectionist tendencies and international trade, there are still several unknowns concerning how this will unfold. There are severe concerns about the collapse of American leadership in trade matters, following its isolationist tendencies (Milanovic, 2019; Albertoni and Wise, 2021). Global alliances have become more fragile and bilateral agreements have taken precedence over multilateralism recently, which does not favor global trade but corporativism. One of the major unknowns relates to the role of the WTO as arbiter of international trade in the near future, given its informal abdication from its functions following the blocking of its judiciary structure by the Trump Administration (Johnson, 2019). The WTO needs stronger and better leadership, meaning more leaders convinced about the WTO’s goals regarding multilateralism and the promotion of greater free international trade. This would enable the WTO to recover from the damage done to cross-border trade networks and global value chains, by, for example, facilitating newer free trade agreements or by the direct elimination of certain customs duties. Risks of harsher unilateralism pose a serious threat to the future growth of trade volumes after their recovery, which is strongly underway. Dynamic trade structures need to persist, and private agency and initiative, supported by institutional structures, should open avenues for innovation and economic development.
The destructive duo formed by nationalism and protectionism are another pandemic that menaces our economies and societies. We still have time to prevent massive contagion .
6. Endnotes
1. For a summary of the relevant literature regarding empirical studies of protectionism see: Takacs (1981), Kleain (1984), Garmann (2014), Osabouhien et al. (2014), Barattieri et al. (2021).
7. References
Albertoni, N. and Wise, C., 2021, International Trade Norms in the Age of Covid-19 Nationalism on the Rise? Fudan J. Hum. Soc. Sci. 14, 41–66.
Grossman, Gene M., and Kenneth Rogoff, 1995, Handbook of International Economics, Volume III (Amsterdam: Elsevier Science Publishers B.V.).
Javorcik, B ,2020, Global supply chains will not be the same in the post-COVID-19 world, Chapter 8 in R Baldwin and S Evenett (eds.), COVID-19 and Trade Policy: Why Turning Inward Won’t Work, a VoxEU.org eBook, CEPR Press.
Johnson, Keith, 2019, How Trump May Finally Kill the WTO, Foreign Policy
Jordà, Oscar, 2005, Estimation and Inference of Impulse Responses by Local Projections, American Economic Review, vol. 95(1), pp. 161-182.
Milanovic, Branko, 2019, Capitalism, Alone: The Future of the System That Rules the World, Harvard University Press.
Smith, Adam, 1776, An Inquiry into the Nature and Causes of the Wealth of Nations, W. Strahan and T. Cadell.
Topalova, Petia, and Amit Khandelwal, 2011, Trade Liberalization and Firm Productivity: The Case of India, The Review of Economics and Statistics, vol. 93(3), pp. 995-1009.
World Bank database, 2021.
IMF, World Economic Outlook Database, 2021.
WTO database, 2021. Note: All graphs and tables are author’s own work, produced with data from World Bank, IMF and WTO databases.
